6.2.2 Risk management and news
News-based analytics are useful in risk management. The difference in return distribu-
tions associated with news have been used to forecast changes in beta, covariances, and
volatility.
In ‘‘Does beta move with news? Firm-specific information flows and learning about
profitability,’’ Patton and Verardo (2009) show that beta (systematic risk) of individual
stocks increases by an ‘‘economically and statistically significant amount on days of
firm-specific news announcements, and reverts to its average level two to five days later.’’
They find marked differences across sectors, an effect we also observed in the research
described in this chapter.
In ‘‘Practical issues in forecasting volatility,’’ Ser-Huang Poon and Nobel Economics
Laureate Clive Granger (2005) describe the combination of GARCH and option implied
volatilities and concluded that the best and most elaborate quantitative models did not
rival predictions based on implied volatilities. In their conclusion, they write, ‘‘a poten-
tially useful area for future research is whether forecasting can be enhanced by using
exogenous variables’’ such as news.
These ideas are extended by Mitra, Mitra, and diBartolomeo (2008) who use implied
volatilities for stocks with options, and exogenous information from news to improve a
multifactor model of equity risk, addressing the issue that ‘‘traditional factor models fail
to update quickly as market conditions change. It is desirable that the risk model
updates to incorporate new information as it becomes available and...introduce a
factor model that uses option implied volatility to improve estimates of the future
covariance matrix. We extend this work to use...quantified news...to improve risk
estimates as the market sentiment and environment changes.’’
6.2.3 Broad long-period analysis of the relation between news and stock returns
In a study first published in 2006, Tetlock, Saar-Tsechansky, and Macskassy looked at
more than 350,000 news stories about S&P 500 companies that appeared in theWall
Street Journaland on the Dow Jones News Service from 1984 to 2004. They used a
massive program called the General Inquirer to gauge the sentiment of these stories. The
General Inquirer is the result of over 20 years of research sponsored by the US National
Science Foundation and the British and Australian National Research Councils. It
started out as PL/I programs running on IBM mainframes in the 1980s. The current
version is hosted (somewhat sporadically lately) on the web by Harvard’s Psychology
Department and is available for anyone to use there. It has spawned dozens of PhD
dissertations, many of which have added language profiles to characterize the sentiment
and content of text.
Tetlock, Saar-Tsechansky, and Macskassy scored those 350,000 stories, containing
over 100 million words, for positive or negative sentiment using the General Inquirer,
and summarized the results in an event study chart showing abnormal returns to stocks
with positive and negative stories. It is shown in Figure 6.1.
These event studies aggregate the results over 20 years (1984–2004). The vertical line
in the center of the chart indicates the date the story appeared. The sentiment measures
appear to work very well. Positive sentiment lines all go up and negative sentiment lines
all go down. But also notice there’s a huge amount of what first appears to be pre-event
Relating news analytics to stock returns 151